Logistic and energy substitution models for electricity forecasting: A comparison using New Zealand consumption data

Abstract The potential of using simple logistic curves for forecasting electricity consumption is investigated with reference to New Zealand sectoral data. In Part A of the paper, fitting of logistic trend curves to historical consumption data uses a Fibonacci search technique to establish optimal asymptotes. The results obtained indicate that current electricity consumptions are approaching saturation. In Part B, a description of an Energy Substitution Model and its assumptions is given. Generally, the Energy Substitution Model is capable of modeling the substitution process among various secondary energy forms, namely, coal, oil, electricity, and gas. The Energy Substitution Model predicts a decline of electricity consumption in New Zealand, a future different from that predicted by the Logistic Model of Part A. This is seen as being due to the unstable penetration rate of natural gas causing significant uncertainties in the results of the former.